My perspective on this (or more generally on the question of whether the future is likely to involve realizing a large fraction of the possible value it could have, whatever form it turns out "value" takes) is perhaps a bit more hopeful. In my view, the question only makes sense if we are moral realists. If there are no objective facts about morality, then I don't see why we should care whether our own preferences or someone else's win out. Furthermore, I think worrying about these questions is probably pointless unless two other things are true: that we have some way of discovering moral facts and that those discoveries have some way of influencing our actions. Unless those two are somehow true, we have no reason to think our efforts can in expectation increase the amount of value realized in the world. 

So far this is a somewhat pessimistic take, but I'm optimistic in a world where all three of these conditions are true, which in some sense is (IMO) the only world where this conversation makes any sense to have. In that world, we should expect that increasing the amount of things like intelligence, time to devote to research/reflection, and focus on studying moral questions in expectation leads toward getting closer to the true morality. Welfareans (or more broadly whatever target will produce the most true value) may indeed get enough advocates just by virtue of society making more progress on these moral questions. As an example, society today includes lots of advocates for groups like women, LGBT people, people of color etc., when historically the only advocates were a "tiny subset of crazy people." But of course moral progress is at best an extremely messy and incremental - factory-farmed animals are the victims of lots of people being either indifferent or wanting to maximize something other than welfare (profit, tasty food for humans etc.), and the impact of animal advocates has not been sufficient to prevent a massive explosion of suffering. 

Still, on net I lean towards thinking that given the opportunity for study and reflection (and given the three conditions described above), we can be optimistic that we will drive toward the things that matter. Therefore, focusing on the efforts to prevent existential catastrophe or value lock-in may be among the best things we could do to ensure that we're not leaving a huge fraction of the possible value of the future on the table. That may be easier said than done, since preventing value lock-in in practice means preventing people with maximizing ideologies from successfully carrying out that maximization at least for some period of time. But that makes me hopeful that existing EA efforts may not be too far off the mark. 

I really enjoyed this! Very important crux for how well the future goes. You may be interested to know that Nick Bostrom talks about this, he calls them super-beneficiaries.

if the value of welfare scales something-like-linearly

I think this is a critically underappreciated crux! Even accepting the other parts, it's far from obvious that the intuitive approach of scaling value linearly in the near-term and locally is indefinitely correct far out-of-distribution; simulating the same wonderful experience a billion times certainly isn't a billion times greater than simulating it once..

I take your point about "Welfareans" vs hedonium as beings rather than things, perhaps that would improve consensus building on this. 

That being said, I don't really expect whatever these entities are to be anything like what we are accustomed to calling persons. A big part of this is that I don't see any reason for experiences to be changing over time; they wouldn't need to be aging or learning or growing satiated or accustomed. 

Perhaps this is just my hedonist bias coming through -  certainly there's room for compromise. But unfortunately my experience is that lots of people are strongly compelled by experience machine arguments and are unwilling to make the slightest concession to the hedonist position. 

Changed my mind, I like this. I'm going to call them Welfareans from now on.

I haven't thought too deeply about it, and it would be a convenient outcome which makes me skeptical, but it seems plausible to me that as technology improves (e.g. gene editing, brain emulation), humans would try to apply it to become more like Welfarians. At least personally, if there was some relatively safe way to change my neurochemistry to derive x% more pleasure from pleasurable experiences, all else equal, I'm confident I'd seriously consider the option. 

I would assume the default case is that the tech would improve total hedonistic or preference satisfaction value for any given individual. Therefore my statement only holds in the cases where credence is given to these value types. Even if moral realism is true, and hedonism/PF were the only contributing values, I'd still guess some deliberate steering would be necessary to reach Welfarian status and not fall short and only be considered a... Hufarian maybe?

We should become our own version of Welfareans at our own pace living peacefully with Welfareans - towards win-win theory

Your Welfareans sound a lot like Nozick's utility monster to me. Do you agree to that comparison?

I hope you have a billion years to evolve your Welfareans. Since humans already have that behind us and our evolution has been going pretty good and faster in the last 200 years maybe in a hundred or so we’ll be getting pretty close to your imagination of Welfareans. 

My thesis is that the core driving bit of humans is love, but that most of what love is, has not fully unfolded yet, like a flower not yet fully blossomed. So we just keep pushing forward, especially on benefitting others (like EA) and the unfolding will continue.